Long circulating RNAs identified as biomarker candidates for risk stratification in childhood acute lymphoblastic leukemia

Analysis of tumoral RNA is essential for childhood acute lymphoblastic leukemia (ALL) diagnosis and prognosis to detect mutations, cryptic fusions and gene expression patterns. Small vesicles circulating in the blood, such as exosomes carry nucleic acids that might be involved in cancer formation and progression, thus presenting a potential for discovery of non-invasive biomarker for diagnosis and monitoring. Here we report the identification of 47 circulating exosome-encapsulated RNA transcripts (30 mRNAs, 5 lncRNAs, 9 circRNAs and 3 pseudogenes) constituting putative biomarker in 2 molecular subgroups of childhood ALL patients. Exosomes were purified from peripheral blood collected at diagnosis from 10 patients positive for the ETV6-RUNX1 and TCF3-PBX1 fusions as well as from conditioned culture medium (CCM) of their representative cell models. Long transcriptome profiling via total RNA sequencing, supported by the implementation of a novel data analysis pipeline confirmed the presence of mRNAs, fusion transcripts, long non-coding and circular RNAs in exosomes. Differential expression analysis comparing tumoral and exosomal RNA revealed 2 transcripts enriched in both leukemic subtypes while 38 were exclusively enriched in ETV6-RUNX1+ patients and 7 were exclusively enriched in TCF3-PBX1+ patients. These are the first results highlighting specific enrichment of long RNA transcripts in exosomes among two distinct ALL subtypes, suggesting the need for investigation of larger cohorts to identify and validate novel circulating biomarkers. Overall design: We assembled a cohort of 24 B-ALL patients equally distributed into cases carrying the ETV6-RUNX1 or the TCF3-PBX1 fusion genes. From this cohort, a subset of 10 patients (5 ETV6-RUNX1+ and 5 TCF3-PBX1+) had both total tumoral RNA and total PB exosomal RNA sequenced. As controls for tumoral samples we included 3 B-cell samples CD10+/CD19+ blood cord cells. As controls for exosomes samples we included previously published blood exosome RNAseq datasets of 32 healthy controls (GSE100206). Raw counts and TPM values are available in the processed data under column names Exos_PB_CTRL_SRR5712482 to Exos_PB_CTRL_SRR5712513 For comparison purposes we included previously published blood exosome RNAseq datasets of 12 colorectal carcinoma (CRC) patients (GSE100063), 21 hepatocellular carcinoma patients (HCC) (GSE100207) and 14 pancreatic carcinoma (PAAD) patients (GSE100232). Raw counts and TPM values are available in the processed data under column names Exos_PB_CRC_SRR5687235 to Exos_PB_CRC_SRR5687246, Exos_PB_HCC_SRR5712516 to Exos_PB_HCC_SRR5712536 and Exos_PB_PAAD_SRR5714908 to Exos_PB_PAAD_SRR5714921 Cellular models of ETV6-RUNX1+ B-ALL (REH) and TCF3-PBX1+ B-ALL (697) were also used for protocol optimization and transcriptome analysis. For both cell lines, we sequenced total RNA from whole cell extracts (Cells), microvesicles (MVs), small extracellular vesicles (Sevs) and immuno-affinity purified (CD9/CD63/CD81) exosomes (Exos).